Chapter 2

DATABASE SYSTEM
ARCHITECTURE
 Objective
• to briefly describe the evolution of database
systems and to understand some of the main
data models used and now in use.
• To understand the database system
Architecture
• Estimated time 4 hours
 Outline
• Basic Database Terminologies
• Overview of data models
• Schemas and Instances
• DBMS Architecture
• Data Independence
 Basic Database Terminologies
• Enterprise - an organization : A library, a bank, a university, etc.
• Entity - person, place, thing, or event
– An "object" in the real world that we are interested in like
Student, Customer, Book, …
• Attribute (Field) - a characteristics of an Entity, that has a specific
meaning. Eg. Name, age, telephone, grade, sex, etc.
• Record - a logically connected set of one or more Attributes that
describe an Entity
• File - a collection of related records. For example, a file might
contain data about customers; or students of a certain department in a
university
• Database - collection of Files
 Data Model
• A set of primitives for defining the structure of a
database.
• Database Model is a collection of tools or concepts
for describing:
– Data
– Data relationships
– Data semantics
– Data constraint
• The main purpose of database model is to represent
the data in an understandable way.
 Data Model

• Categories of database models include:

– Object-based
– Record-based
– Physical
Evolution of Major Data Models
 File-Processing Systems
• The earliest database systems – during the 1960s
• Data was stored in flat text files in the operating system.
• Data was also grouped into records
e.g. a record of Customer data in a bank might look like this:
 Number, Name, FathersName, PhoneNumber, Address,
SavingsAccountNumber, SavingsAccountBalance, CurrentAccountNumber,
CurrentAccountBalance

123,Tesfay, Kebede, 04 407600, Yeka kefle ketema Addis Ababa, 1234, 1000,
9876, 1000

• A file contains data about one entity e.g. Customer is an entity. Each record in the
file represents an instance of that entity.
• Each row is a record; the file represents a record type.
 Cont’d
Disadvantages
• Programmers had to write specific programs to carry out each tasks - so a lot of
work was involved.
– e.g. to retrieve all customer records or to find the customer record for a given customer
number.
• It could also lead to data being duplicated in different files
– e.g. think of a bank’s database. A customer’s name and phone number could appear in a
file containing all savings accounts. If the same customer also has a current account, their
name and phone number would also appear in the file containing all current accounts. Or
if a customer has 2 savings accounts – need 2 rows in the file –name and phone number
are repeated.
•  If the customer changes their phone number – it has to be changed in both files.
•  This system worked, but didn’t allow data to be related – e.g. to have the
customer name and phone number in one place, and relate a savings account to
it and a current account to it.
 Record-based Data Models

• Consist of a number of fixed format records.

• Each record type defines a fixed number of fields
• Each field is typically of a fixed length.
• There are three types of Record-based data models:-
–  Hierarchical Database Model
–   Network Database Model
–  Relational Database Model
Hierarchical Model
 cont’d
• The simplest database model
• Records are represented as nodes or segments
– Nodes are arranged in a hierarchical structure as sort
of upside-down tree. The tree may be of arbitrary depth
• Each record contain multiple fields, where each field
may contain either data values like integer, real, text
or a pointer to a record. The pointer is not allowed to
form a cycle. Some hierarchical DBMS support null
values or variable-length fields .
 Cont’d
•  The top node is the root node
• The relationship between parent and child is one-
to-many and one-to-one
–   A parent node can have more than one child node
–  A child node can only have one parent node
•   Relation is established by creating physical link
between stored records (each is stored with
a predefined access path to other records)
 Cont’d
• The data is still stored in a flat file….like previously…
but now the flat file has 3 different record types in it
and the record types are related:
Customer: 123, Teffaye , Abebe, 04 407600, Yeka Kefle
ketema,
Account: Savings, 123456, 5400
Transaction:…..
Transaction:….
Account: Current,278654, 2000
Customer: ……………
 Advantages
– Hierarchical Model is simple to construct and operate on
– Conceptual simplicity – easy to understand the model layout
– Data independence (a change in a data type will be
automatically cascaded throughout the database by the DBMS,
thereby eliminating the need to make changes in the program
segments that reference the changes data type)
– Database integrity – always a link between parent and child
– Efficiency – very efficient when it contains a large volume of
data in 1:M relationships and whose relationships are fixed over
time
 Cont’d
• Searching of any record in a hierarchical tree is
very fast since the hierarchical databases uses
contiguous storage for hierarchical structures.
• Language is simple; uses constructs like GET,
GET UNIQUE, GET NEXT, GET NEXTWITHIN
PARENT etc.

Fundamentals Database Systems 17

 Disadvantage
– Complex implementation – detailed knowledge of the physical data
storage characteristics is required by the designers and programmers
– Difficult to manage – relocation of segments requires application
changes
– Lacks structural independence
– Complex applications programming and use – programmers and end
users must know precisely how the data are physically distributed
within the database
– Lack of standards – no standard DDL and no DML
 Cont’d
•  Navigational and procedural nature of processing.
• A programmer still has to write the program to
access the account information, but now the
program navigates through the hierarchy e.g. to
find the balance of the savings account for a given
customer number.
• does not support much consistency and security

Fundamentals Database Systems 19

 Cont’d
• Implementation limitations – difficult to support M:N
relationships
– a child node in the tree cannot have more than one
parent. So if the same account is associated with 2
customers (e.g. a husband and wife have a joint account
and also separate accounts), we cannot link the one
account record to 2 different customer records. But we
can have the account record appearing in two tree
branches. This can lead to duplicated data – and
inconsistent data, if the account is not updated in all the
branches.

Fundamentals Database Systems 20

 Cont’d
• To add new record type or relationship, the
database must be redefined and then stored
in a new form.
 Network Model
• also usesrecords and links but in a different way.
• Instead of putting data in a hierarchy, it structures
data in a network

Customer, Account and Transaction record types.

 Cont’d
• Like hierarchical model network model is a collection of physically
linked records.
• Collection of records in 1:M relationships
– Relationships are modelled using sets.
• In a set, there is one owner-record type--Equivalent to the hierarchical model’s parent
eg. (Customer) and 1 or more member record types Equivalent to the hierarchical
model’s child eg.(Account).This is a set type, let us name it CustomerAccount. It models
the relationship between Customer & Account – that a Customer can have 1 or more
accounts. There will be many occurrences of the CustomerAccount set in the database
– one for each Customer. A set occurrence relates one record from the owner record
type (Customer) to the set of records from the member record type related to it
(Account).For each CustomerAccount set, there is a network of links from the Customer
record to the Account records.
• A record can appear as a member in more than one set i.e., a member may have multiple
owners
• Can you identify another set type in this model?
 Cont’d
• The data is still stored in files but now we also
have to define the set types.
• The entry points are records that can be
searched. An entry point is implemented as an
index on a list of records – so the index must be
created before such queries can be run against
the database.
• In terms of data structure can you think of how
entry implemented ?
Registrar system

Fundamentals Database Systems 25

 Advantages

– Conceptual simplicity
– Data access flexibility – no need for a preorder traversal
– Promotes database integrity – must first define the owner
and then the member record
– Data independence
– Network Model is able to model complex relationships and r
epresents semantics of add or delete on the relationships.
– Can handle most situations for modeling using record types
and relationship types.
– Conformance to standards

Fundamentals Database Systems 26

 Disadvantages

• System complexity
• Lack of structural independence
•  Navigational and procedural nature of processing.
• Language is navigational; uses constructs like FIND, FIND
member, FIND owner, FIND NEXT within set, GET etc.
Programmers can do optimal navigation through the database
• Database contains a complex array of pointers
that thread through a set of records.
• Little scope for automated "query optimization”
 Relational Model
• Developed by Dr. Edgar Frank Codd in 1970 (famous paper, 'A
Relational Model for Large Shared Data Banks').
• Terminologies originates from the branch of mathematics called set
theory and relational algebra.
• Can define more flexible and complex relationship.
•  Viewed as a collection of tables called “Relations” equivalent to
collection of record types.
• Relation: Two dimensional table.
•  Stores information or data in the form of tables  rows and
columns.
– A row of the table is called tuple  equivalent to record.
–  A column of a table is called attribute  equivalent to fields.

Fundamentals Database Systems 28

 Cont’d
•  Data value is the value of the Attribute.
•  Records are related by the data stored jointly in the fields
of records in two tables or files. The related tables contain
information that creates the relation.
•  No physical consideration of the storage is required by the
user.
•  Many tables are merged together to come up with a new
virtual view of the relationship.
• The relational data model is implemented through a very
sophisticated relational database management system
(RDBMS).
Cont’d
 Cont’d
• The RDBMS manages all of the physical details, while the user sees the
relational database as a collection of tables in which data are stored.
• The user can manipulate and query the data in a way that seems
intuitive and logical.
• Conducts searches by using data in specified columns of one table to
find additional data in another table.
• In conducting searches, a relational database model matches
information from a field in one table with information in a
corresponding field of another table to produce a third table that
combines requested data from both tables.

Fundamentals Database Systems 32

 Advantages
• Structural independence – changes in the
relational data structure do not affect the DBMS’s
data access in any way
• Improved conceptual simplicity by concentrating
on the logical view
• Easier database design, implementation,
management, and use
• Ad hoc query capability - SQL
• Powerful database management system
Fundamentals Database Systems 33
 Disadvantages
• Substantial hardware and system software
overhead
• Can facilitate poor design and implementation

Fundamentals Database Systems 34

 Object relational
• An Object relational model is a combination of a
Object oriented database model and a Relational
database model. So, it supports objects, classes,
inheritance etc. just like Object Oriented models and
has support for data types, tabular structures etc. like
Relational data model.
• One of the major goals of Object relational data model
is to close the gap between relational databases and
the object oriented practices frequently used in many
programming languages such as C++, C#, Java etc.
 Advantages
• Inheritance
– The Object Relational data model allows its users to inherit objects,
tables etc. so that they can extend their functionality. Inherited objects
contains new attributes as well as the attributes that were inherited. 
• Complex Data Types
– Complex data types can be formed using existing data types. This is
useful in Object relational data model as complex data types allow
better manipulation of the data.
• Extensibility
– The functionality of the system can be extended in Object relational
data model. This can be achieved using complex data types as well as
advanced concepts of object oriented model such as inheritance.
 Disadvantages
The object relational data model can get quite
complicated and difficult to handle at times as
it is a combination of the Object oriented data
model and Relational data model and utilizes
the functionalities of both of them.
 Deductive Model

• In a deductive db system, rules can be defined – the

rules deduce or infer additional information from the
facts stored in the database.
• Deductive databases are really a type of knowledge
base, used in the area of AI (Artificial Intelligence).
• A deductive db has facts and rules in it. Facts are
stored similar to relations in a relational database, but
attribute names are not necessary.
• Rules are specifications that can be applied to the facts
– to produce new information.
 Cont’d
• The rules are defined using a declarative language
(what, rather than how…).
• The system has an inference engine that deduces new
facts from the db by interpreting the rules.
• The deductive model is closely related to the relational
model; it also has its basis in a branch of mathematics
(domain relational calculus).
•  In a deductive database system, the emphasis is on
deriving new knowledge from existing data by
supplying rules based on knowledge of the real world.
Summary of Data Model
 Schema & Instance
There are two basic components of the
database: the definition of the Relation
(Schema) and the actual data stored in each
table (Instance)
 Database Schema
• Describes how data is to be structured, defined at
setup/design time (also called "metadata“)
• it specifies name of relation, the name of each field and
the type of each field
• Example:
– Students( sid: string, name: string, login:
string, age: integer, gpa: real)
• The preceding schema says that each record in the
Students relation has five fields, with field names and
types as indicated
 Database Schema

•  The structure of a database that

– Captures data types, relationships and constraints
in data
• Is independent of any application program
• Changes infrequently
 Database Instance or State:
• is the collection of the actual data contained in the
database at a particular point of time
• Also called State or Snap Shot or Extension of the
database
• Since Instance is actual data of database at some
point in time, it changes rapidly
• To define a new database, we specify its database
schema to the DBMS (database is empty), database is
initialized when we first load it with data
 Three levels ANSI_SPARC Architecture

• The Systems Planning and Requirements

Committee of the American National
Standards Institute encapsulated the concept
of schema in its three-level database
architecture model, known as the ANSI/SPARC
architecture, which is shown in the diagram
below
Three-level ANSI-SPARC Architecture of a Database System
 ANSI/SPARC three-level architecture
• ANSI = American National Standards Institute
• ANSI/X3 = Committee on Computers and
Information Processing
• SPARC = Standards Planning and Requirements
Committee
• The ANSI/SPARC model is a three-level database
architecture with a hierarchy of levels, from the
users and their applications at the top, down to
the physical storage of data at the bottom.
• The Architecture of most of commercial DBMS
are available today is mostly based on this
ANSI-SPARC database architecture
• Proposed to support DBMS characteristics of:
– Program-data independence
– Support of multiple views of the data
 Objective of Three-level ANSI_SPARC architecture

• It allows independent customized user views:

– Each user should be able to access the same data, but have a
different customized view of the data
• These should be independent: changes to one view should not affect others
• It hides the physical storage details from users:
– Users should not have to deal with physical database storage details
• Database Administrator should be able to change conceptual
structure of database without affecting all users
• Database Administrator should be able to change database
storage structures without affecting the users’ views
• External Level
– Users’ view of the
database
• Conceptual Level
– Community view of
the database
• Internal Level
– Physical
representation of the
database on the
computer
 Cont’d
External Level
– The external or view level includes a number of external
schemas or user views.
– Each external schema describes the part of the database
that a particular user group is interested in and hides the
rest of the database from that user group.
– It excludes irrelevant data as well as data which the user
is not authorized to access.
– Implemented using a representational data mode
 Cont’d
Conceptual Level
– The conceptual level is a way of describing what data is stored
within the whole database and how the data is inter-related
– The conceptual level has a conceptual schema, which
describes the structure of the whole database for a
community of users.
– concentrates on describing entities, data types, relationships,
user operations, and constraints.
– Usually, a representational data model is used to describe the
conceptual schema when a database system is implemented.
– This level does not specify how the data is physically stored
 Cont’d
Internal Level
• The internal level has an internal schema, which
describes the physical storage structure of the
database.
• The internal level involves how the database is
physically represented on the computer system
• It describes how the data is actually stored in the
database and on the computer hardware
• The internal schema uses a physical data model
Three levels ANSI_SPARC Architecture
Schema Example
 Three levels ANSI_SPARC Architecture
Schema Example
• External Schema (View):
– Course_info(cid:string,sid:string,name:string)
• Conceptual (logical) schema:
– Students(sid: string, name: string, login: string, age:
integer, gpa:real)
– Courses(cid: string, cname:string, credits:integer)
– Enrolled(sid:string, cid:string, grade:string)
• Internal(Physical) schema:
– Relations stored as unordered files
– Index on first column of Students
 Database schemas

• There are three different types of schema corresponding to

the three levels in the ANSI-SPARC architecture
• The external schemas
– describe the different external views of the data and there may be
many external schemas for a given database
• The conceptual schema
– describes all the data items and relationships between them of the
whole database
– There is only one conceptual schema per database
• The Internal schema
– describes physical storage structures and access paths
– There is only one internal schema per database
• The overall description of a database is called the database
schema
 Data Independence
• A very important advantage of using a DBMS is
that it offers Data Independence
• That is, application programs are insulated from
changes in the way the data is structured and
stored
• Data independence is achieved through use of
the three levels of data abstraction (i.e., External
level,Conceptual level & Internal level)
 Logical Data Independence
• The capacity to change the conceptual schema without
having to change the external schemas and their application
programs
• Conceptual schema changes e.g. addition/removal of
entities should not require changes to external schema or
rewrites of application programs
• Here are some examples of changes in the logical layer that
can be safely made thanks to logical data independence:
– Adding a new database object
– Adding data items to an existing object
 Physical Data Independence
• The ability to modify the internal schema without
changing the conceptual schema
• The conceptual schema hides details such as how
the data is actually laid out on disk, the file
structure, and the choice of indexes
• Physical schema changes e.g. using different file
organizations, storage structures/devices should not
require change to conceptual or external schemas
Data Independence and the ANSISPARC
Three-Level Architecture
 Two-Tier Client-Server Architecture

• Client manages main business

and data processing logic and
user interface
• Server manages and controls
access to database
• Disadvantages:
– ‘Fat’ client, requiring
considerable resources on
client’s computer to run
effectively
– Significant client side
administration overhead
 Three-Tier Client-Server Architecture

• a type of multi-tier computing architecture

in which an entire
application is distributed across three
different computing
layers or tiers
– User interface layer – runs on client
– Business logic and data processing layer –
middle tier runs on a server
(application server)
– DBMS – stores data required by the middle
tier. This tier may be on a
separate server (database server)
• Advantages:
‘Thin’ client, requiring less expensive
hardware Application maintenance
centralized Easier to modify or replace one
tier without affecting others